Feedback Optimizing MPC for Load Frequency Control and Economic Dispatch

Abstract

With increasing renewable generation, demand response, and deregulation, power networks are becoming more uncertain, time-varying, and strongly coupled. As a result, the conventional approach of performing separate economic dispatch (ED) and load-frequency control (LFC) operations may no longer guarantee smooth and cost-efficient regulation of frequency across interconnected power networks. To address this, we present a tracking model predictive control (MPC) algorithm which simultaneously achieves economic dispatch and secondary frequency control in a multi-area power network. A unique feature of the proposed algorithm is that it exploits the implicit feedback in MPC to regulate the interconnected power system towards steady-state equilibria that solve a multi-area economic dispatch problem, without explicitly computing the latter as a reference to be followed or estimating the unknown disturbances. This feedback-based optimization approach endows the algorithm with inherent robustness to uncertainty (such as unknown step changes in the demand). Simulation results for a two-area power network show improved steady-state economic performance compared to standard MPC-based frequency control schemes, and better dynamic performance compared to other feedback-based optimization schemes.